David A. Wheeler measures the size of a contemporary GNU/Linux distribution by analyzing Red Hat Linux 7.1 and estimating what it would cost to recreate it using conventional proprietary development. The distribution contains over 30 million physical source lines of code (SLOC) and, using the COCOMO model in year-2000 U.S. dollars, represents roughly 8,000 person-years of effort and over $1 billion in development cost. That is a rise of more than 60% from Red Hat 6.2 (about 17+ million SLOC, ~4,500 person-years, ~$600 million) accumulated over roughly 13 months. The largest components by size are the Linux kernel (including device drivers), Mozilla, X Window System, gcc, gdb, core binary tools, Emacs, LAPACK, GIMP, and MySQL. Language composition is dominated by C (71%) and C++ (15%), followed by shell, Lisp, assembly, Perl, Fortran, Python and others.
The methodology counts physical SLOC (non-blank, non-comment lines), excludes duplicate and automatically generated files, ignores non-x86-only packages, and uses md5 checksums and heuristics to classify files and languages. Licensing analysis shows copyleft licenses dominate: slightly over half the code uses the GPL and GPL/LGPL together account for 63% of the distribution, while public-domain code is about 0.2%. Red Hat 7.1 was chosen as representative due to its market share and the fact many other distributions reuse similar components.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.